The analog dynamic memory circuit (hereinbelow called ADM) has been known as a memory for memorizing an analog signal waveform by circulating an analog signal waveform in a closed loop comprising a delay element such as a CCD. During this circulation, the amplitudes of the analog signal waveform are maintained by setting the gain of the loop as 1. Comparison of two analog signal waveforms in one of the applications in which ADMs are used. By means of this comparison, it is determined whether or not the analog signal waveforms are identical.
This comparison can be attained, for example, by comparing the differences between the amplitudes of the analog signal waveforms. When there are no differences, the analog signal waveforms are regarded to have the same waveform. It is assumed that two analog signals are respectively represented by S.sub.1 (t) and S.sub.2 (t) as follows: EQU S.sub.1 (t)=A.sub.1 .multidot.sin (2.pi.f.sub.1 t) (1) EQU S.sub.2 (t)=A.sub.2 .multidot.sin (2.pi.f.sub.2 t) (2)
When the above two signals fulfill the following requirements, that is A.sub.1 =A.sub.2 in amplitude and f.sub.1 =f.sub.2 in frequency(pitch), these signals are regarded as being identical. In other words, there is no difference between the two signals. EQU D=S.sub.1 (t)-S.sub.2 (t)=0 (3)
In general, the analog signals stored in two ADMs are different in amplitude and frequency(pitch), and thus a meaningless composite signal is produced if these differences exist. For instance, taking voice recognition into consideration, a word spoken by a certain person is stored in a first ADM, then the same word spoken by the same person is stored in a second ADM. But the audio signal stored in the first ADM and the signal stored in the second ADM do not always have the same amplitudes and frequencies therebetween. Even if the same person speaks the same word, the magnitudes and the tones of the voice are sometimes different. If the audio signals are represented by simple sine functions, the process of the voice recognition can also be represented as follows: ##EQU1##
Therefore, when the magnitudes and tones of the voice for the same word of the same person are compared by using the ADMs, it is required that the amplitudes and frequencies of the audio signals stored in the two ADMs be identical.